AI Infrastructure Engineer (LLM Training & GPU Inference) — Berlin, Germany

The Opportunity

Fin — the AI Customer Agent company behind Intercom's highest-performing AI products — is looking for Senior+ AI Infrastructure Engineers to build the systems that train and serve Fin's next generation of AI products. Fin builds from the GPU all the way up to a user agent that resolves millions of customer service queries every month, and this Berlin-based team sits at the cutting edge of modern AI infrastructure.

What You'll Do

• Build and scale the training pipelines and inference systems behind custom models like Fin Apex — which outperforms frontier models on customer service tasks
• Work on model training and model inference at scale, including low-level GPU coding (CUDA, Triton)
• Join a small, highly technical team that owns the full AI stack — from GPUs to user-facing agents
• Push the performance, reliability and cost-efficiency of AI infrastructure used by nearly 30,000 global businesses

What We're Looking For

• A track record working on model training or model inference at scale
• Experience with low-level GPU programming (CUDA, Triton) — one area is great, multiple is better
• Deep systems thinking and a builder's mindset
• Comfort working across the full AI infrastructure stack

Why Intercom / Fin

You'll work on genuinely frontier AI infrastructure — custom models that beat frontier baselines in production — in one of the largest private software companies in the world. Small team, massive impact, and the chance to define how AI customer agents are built.

Ready to build the GPU-to-agent stack? Apply below.

---

🚀 Ready to Apply?

👉 Click Here to Apply Directly for this Position